Long Duration Persistent Photocurrent in 3 nm Thin Doped Indium Oxide for Integrated Light Sensing and In-Sensor Neuromorphic Computation

摘要

Miniaturization and energy consumption by computational systems remainmajor challenges to address. Optoelectronics based synaptic and lightsensing provide an exciting platform for neuromorphic processing and visionapplications offering several advantages. It is highly desirable to achievesingle-element image sensors that allow reception of information andexecution of in-memory computing processes while maintaining memory formuch longer durations without the need for frequent electrical or opticalrehearsals. In this work, ultra-thin (<3 nm) doped indium oxide (In_2O_3) layersare engineered to demonstrate a monolithic two-terminal ultraviolet(UV) sensing and processing system with long optical state retentionoperating at 50 mV. This endows features of several conductance stateswithin the persistent photocurrent window that are harnessed to showlearning capabilities and significantly reduce the number of rehearsals. Theatomically thin sheets are implemented as a focal plane array (FPA) for UVspectrum based proof-of-concept vision system capable of pattern recognitionand memorization required for imaging and detection applications. Thisintegrated light sensing and memory system is deployed to illustratecapabilities for real-time, in-sensor memorization, and recognition tasks. Thisstudy provides an important template to engineer miniaturized and lowoperating voltage neuromorphic platforms across the light spectrum based onapplication demand.